linux-kernel - [tip:perf/core] kprobes/x86: Do not jump-optimize kprobes on irq entry code

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <tip-d9f5f32a7d17f4906a21ad59589853639a1328a0@git.kernel.org>
Date:   Thu, 10 Aug 2017 09:36:47 -0700
From:   tip-bot for Masami Hiramatsu <tipbot@...or.com>
To:     linux-tip-commits@...r.kernel.org
Cc:     linux-kernel@...r.kernel.org, mhiramat@...nel.org,
        tglx@...utronix.de, davem@...emloft.net, jcmvbkbc@...il.com,
        hpa@...or.com, starvik@...s.com, torvalds@...ux-foundation.org,
        anil.s.keshavamurthy@...el.com, jesper.nilsson@...s.com,
        ananth@...ibm.com, francis.deslauriers@...icios.com,
        mingo@...nel.org, peterz@...radead.org, chris@...kel.net,
        ysato@...rs.sourceforge.jp
Subject: [tip:perf/core] kprobes/x86: Do not jump-optimize kprobes on irq
 entry code

Commit-ID:  d9f5f32a7d17f4906a21ad59589853639a1328a0
Gitweb:     http://git.kernel.org/tip/d9f5f32a7d17f4906a21ad59589853639a1328a0
Author:     Masami Hiramatsu <mhiramat@...nel.org>
AuthorDate: Thu, 3 Aug 2017 11:39:26 +0900
Committer:  Ingo Molnar <mingo@...nel.org>
CommitDate: Thu, 10 Aug 2017 16:28:53 +0200

kprobes/x86: Do not jump-optimize kprobes on irq entry code

Since the kernel segment registers are not prepared at the
entry of irq-entry code, if a kprobe on such code is
jump-optimized, accessing per-CPU variables may cause a
kernel panic.

However, if the kprobe is not optimized, it triggers an int3
exception and sets segment registers correctly.

With this patch we check the probe-address and if it is in the
irq-entry code, it prohibits optimizing such kprobes.

This means we can continue probing such interrupt handlers by kprobes
but it is not optimized anymore.

Reported-by: Francis Deslauriers <francis.deslauriers@...icios.com>
Tested-by: Francis Deslauriers <francis.deslauriers@...icios.com>
Signed-off-by: Masami Hiramatsu <mhiramat@...nel.org>
Cc: Ananth N Mavinakayanahalli <ananth@...ibm.com>
Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@...el.com>
Cc: Chris Zankel <chris@...kel.net>
Cc: David S . Miller <davem@...emloft.net>
Cc: Jesper Nilsson <jesper.nilsson@...s.com>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Max Filippov <jcmvbkbc@...il.com>
Cc: Mikael Starvik <starvik@...s.com>
Cc: Peter Zijlstra <peterz@...radead.org>
Cc: Thomas Gleixner <tglx@...utronix.de>
Cc: Yoshinori Sato <ysato@...rs.sourceforge.jp>
Cc: linux-arch@...r.kernel.org
Cc: linux-cris-kernel@...s.com
Cc: mathieu.desnoyers@...icios.com
Link: http://lkml.kernel.org/r/150172795654.27216.9824039077047777477.stgit@devbox
Signed-off-by: Ingo Molnar <mingo@...nel.org>
---
 arch/x86/kernel/kprobes/opt.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/arch/x86/kernel/kprobes/opt.c b/arch/x86/kernel/kprobes/opt.c
index 69ea0bc..4f98aad 100644
--- a/arch/x86/kernel/kprobes/opt.c
+++ b/arch/x86/kernel/kprobes/opt.c
@@ -39,6 +39,7 @@
 #include <asm/insn.h>
 #include <asm/debugreg.h>
 #include <asm/set_memory.h>
+#include <asm/sections.h>
 
 #include "common.h"
 
@@ -251,10 +252,12 @@ static int can_optimize(unsigned long paddr)
 
 	/*
 	 * Do not optimize in the entry code due to the unstable
-	 * stack handling.
+	 * stack handling and registers setup.
 	 */
-	if ((paddr >= (unsigned long)__entry_text_start) &&
-	    (paddr <  (unsigned long)__entry_text_end))
+	if (((paddr >= (unsigned long)__entry_text_start) &&
+	     (paddr <  (unsigned long)__entry_text_end)) ||
+	    ((paddr >= (unsigned long)__irqentry_text_start) &&
+	     (paddr <  (unsigned long)__irqentry_text_end)))
 		return 0;
 
 	/* Check there is enough space for a relative jump. */